Optimization landscapes and evolutionary designs
Project Leader: Saman Halgamuge
Staff: Michael Kirley (Computing and Information Systems), Guillermo Narsilio (Infrastructure Engineering)
Collaborators: Toon K Chan (School of Design), Yahui Sun (ANU)
Primary Contact: Saman Halgamuge (firstname.lastname@example.org)
Keywords: artificial intelligence; evolutionary algorithms; optimisation
Disciplines: Mechanical Engineering
Domains: Networks and data in society, Optimisation of resources and infrastructure
As we do not know much about the fitness landscape of the optimisation problem we are trying to solve, it is challenging to pick the right method for a given complex real problem.
Comparisons of algorithms using numerous benchmarks may reveal the better algorithms suitable for the benchmarks. However, we do not know whether the set of benchmarks includes a problem similar to the one we try to solve. The ultimate solution to the algorithm selection problem is the characterisation of the fitness landscape. The solution to this problem becomes even more challenging when the fitness landscape changes dynamically.
We develop models for metaheuristic research which recognises the need to match algorithms to problems. Information theoretic measures as well as semi-empirical approaches to producing mapping from problems to algorithms are investigated. This mapping, if successful, will encapsulate the knowledge gained from the application of metaheuristics to the spectrum of real problems.